Accuracy of Ambulatory Blood Pressure Monitoring
Although ambulatory measurement devices compare favorably with auscultatory methods of blood pressure determination in the clinic, there are substantial differences in average blood pressures obtained during daily life and average blood pressures obtained in the clinic.
Given the dynamic nature of blood pressure, it should not come as a surprise that resting blood pressures in the clinic may not correspond well with blood pressures obtained during an extended period of time comprised of daily activities at work and at home as well as several hours of sleep.
Unusually, the nature of the differences between mean ambulatory and mean clinic blood pressure values often differs between hypertensive and normotensive patients. Among hypertensive patients, clinic-derived measures of blood pressure are typically a good bit higher than average daily blood pressure readings (see Pickering and James, 1989); among normotensive controls, however, differences between clinic and ambulatory blood pressures are often minimal, and the opposite relation has even been reported; that is, ambulatory blood pressures were higher than clinic blood pressures (Cox et al., 1991).
These findings are commonly associated with the speculation that groups of hypertensive patients often comprise a significant portion of isolated clinic hypertensive patients.
As with the lower reference values employed for evaluating home blood pressure measures, there has been considerable discussion regarding the appropriate cutoff criteria for considering a diagnosis of hypertension based upon ambulatory recordings. Since average ambulatory blood pressures are typically lower than clinic-derived measures of blood pressure (Pickering and James, 1989), it has been recommended that ambulatory blood pressures lower than 140/90 mm Hg still be considered within the hypertensive range.
There have been a few different approaches as to how these ambulatory blood pressure diagnostic criteria should be established (see Pickering, 1999b). For example, one approach to this problem is to base these cutoff values upon a certain percentile of distributions of blood pressures of large samples of people. Adopting a 95th percentile criterion for 24-hour blood pressures across studies, however, yields different cutoff points.
While the 95th percentile in a study by O’Brien et al. (1991) was reported to be 134/88 mm Hg, the same percentile from the PAMELA Study revealed criteria of 130/81 mm Hg (Mancia et al., 1995). In a more recent report, Rasmussen et al. (1998) found no differences between ambulatory and clinic measures on a large sample of Danish citizens; had they employed the 95th percentile criteria on the ambulatory blood pressure measurements, their cutoff for considering a diagnosis of essential hypertension would have been 156/90 mm Hg!
Because such disparate criteria have been reported from study to study using this approach, it may not represent the optimal way of handling this issue.
Another strategy for establishing diagnostic criteria for ambulatory blood pressures is to aggregate findings across studies using meta analytic procedures to arrive at uniform cutoff values. Staessen et al. (1991) conducted such a meta-analysis using data on more than 3400 people from 23 studies, revealing a criteria for probable hypertension at 139/87 mm Hg based upon the 95th percentile of the overall sample.
Because the prevalence of hypertension is greater than 5percent of the populations of industrialized countries, the general approach for both of the first two strategies for establishing appropriate criteria for elevated ambulatory blood pressure may miss categorizing many individuals with essential hypertension.
Furthermore, as Pickering (1996b) points out, if hypertensive patients are not included in these studies in an effort to examine mean ambulatory blood pressures for a healthy population, findings regarding ambulatory blood pressures will under-estimate the true population values.
Pickering (1999b) discussed a third strategy for defining normal ambulatory blood pressures involving the calculation of what has been called blood pressure load. Blood pressure load refers to the number of ambulatory blood pressure recordings for a given individual that fall above 140/90 mm Hg during the day and above 120/80 mm Hg at night.
Individuals with sustained essential hypertension exhibit blood pressure loads of 100 percent, and normotensive individuals typically exhibit blood pressure loads of 0 percent. Blood pressure loads exceeding 40percent have been shown to be related to target organ damage in hypertensive patients (White, Dey, and Schulman, 1989). Although this approach shifts the focus from population-based methods for determining risk for hypertension to methods based upon associations with target organ pathology, Pickering questions the arbitrary methods employed for choosing the specific criteria used in determining blood pressure load.
Finally, reference values for ambulatory monitoring could be established by examining the direct relation between ambulatory blood pressures and either measures of target organ pathology or prognosis for cardiovascular disease (Pickering, 1999b). For example, Ohkubo et al. (1998a) followed a rural cohort of over 1500 individuals for six years, during which 117 died. Ambulatory blood pressure monitoring was associated with increased mortality at both high and low levels; increased mortality associated with high ambulatory blood pressures was related to cardiovascular diseases, whereas increased mortality associated with low ambulatory blood pressures was related to other forms of mortality.
Reference values of 134/79 mm Hg for ambulatory 24-hour blood pressure recordings demarcated increased risk for cardiovascular mortality. In this strategy, the cutoff values were determined through association with actual risk for cardiovascular disease complications rather than the selection of a pre-designated percentile of the population.
In summary, there are numerous methods for determining the ambulatory blood pressure level derived from 24-hour recording at which diagnoses of hypertension and associated increased risk for cardiovascular disease should be considered. Acknowledging Pickering’s (1999a) contention that blood pressure should be considered a continuous variable and that efforts at identifying cutoff points for demarcating the moment at which intervention is warranted may be misguided, there is general consensus across these various methods that mean 24- hour blood pressures associated with diagnoses of hypertension and increased risk for cardiovascular disease are lower than the established clinic-derived reference values.
Although many investigators continue to disagree regarding the exact reference values to use in evaluating ambulatory blood pressures, current recommendations indicate that 24- hour ambulatory blood pressures that exceed 135/85 mm Hg should be considered elevated (O’Brien et al., 2000). Additionally, daytime ambulatory blood pressures that exceed 140/90 mm Hg or nighttime ambulatory blood pressures that exceed 125/75 mm Hg should be considered elevated (O’Brien et al., 2000).